[Debian-science-sagemath] RFS for several Sage-related packages

[Doug Torrance]

On Tue, May 26, 2020 at 11:48 AM Dima Pasechnik <dimpase at gmail.com> wrote:
>
>
>
> On Tue, May 26, 2020 at 12:36 PM Doug Torrance <dtorrance at piedmont.edu> wrote:
> >
> > On Tue, May 26, 2020 at 7:13 AM Dima Pasechnik <dimpase at gmail.com> wrote:
> > >
> > > On Tue, May 26, 2020 at 11:29 AM Doug Torrance <dtorrance at piedmont.edu> wrote:
> > > >
> > > > On Tue, May 26, 2020 at 4:15 AM Dima Pasechnik <dimpase at gmail.com> wrote:
> > > > > > Thanks for the link!  It looks out of date with respect to upstream,
> > > > > > though.  The csdp package in Debian works quite well with Macaulay2 in
> > > > > > its current state.  Is there anything in particular that would be
> > > > > > useful to have for Sage in Debian that isn't available right now with
> > > > > > the csdp package?
> > > > >
> > > > > Looking at your package, it seems you are installing things in
> > > > > /usr/local, which is
> > > > > out of line with the usual Debian policy.
> > > >
> > > > That's the upstream default, but the Debian package is patched to
> > > > install everything into /usr.
> > > >
> > > > > That's a bit offtopic, but does M2's csdp package work on macOS?
> > > > > (Sage's csdp does)
> > > >
> > > > I believe it does, as a large proportion of M2's users are Mac users.
> > >
> > > Right - if you only build a static library.
> > > The main reason for autotoolization was an ability to build a dynamic library.
> >
> > That makes sense.  M2 doesn't link against the csdp library, but
> > several of its packages (written in M2's own high-level language)
> > interface with csdp by parsing its output from the command line.
> >
> > Would having a dynamic csdp library in Debian be useful for Sage?
>
> yes, we are building a csdp backend for Sage SDP solver, and
> technologically speaking it needs a dynamic library
> to talk to (same would apply to building Python, and some other languages, extensions, in general).

Sounds good!  I'll see what I can do.